Description_of_observables.md

Description of observables

TheSimulationResult object, provides the following observables (usualy defined as val result=zombieInvasion(...) in the ScalaTask)

Some of the observable are time series. In this case, observable can be aggregated via a temporal step.

The by : Int = 20 notations stand for the temporal step parameter by default. All time-serie indicators take such a by parameter, which samples the original 500 steps time-serie by this fixed step width (for convenience and size of output data).

Observables are:

Agent-related indicators

• humansDynamic(by: Int = 20): Array[Int] sampled time serie (each by time steps) of number of humans
• walkingHumansDynamic(by: Int = 20): Array[Int] sampled time serie of number of walking humans
• runningHumansDynamic(by: Int = 20): Array[Int] sampled time serie of number of running humans
• zombiesDynamic(by: Int = 20): Array[Int] sampled time serie of number of zombies
• walkingZombiesDynamic(by: Int = 20): Array[Int] sampled time serie of number of walking zombies
• runningZombiesDynamic(by: Int = 20): Array[Int] sampled time serie of number of running zombies

Event-related indicators

• def rescuedDynamic(by: Int = 20): Array[Int] sampled time serie of number of rescued humans.
• def killedDynamic(by: Int = 20): Array[Int] sampled time serie of killed zombies
• def zombifiedDynamic(by: Int = 20): Array[Int] sampled time serie of number of zombified humans
• def fleeDynamic(by: Int = 20): Array[Int] sampled time serie of number of humans fleeing from zombies
• def pursueDynamic(by: Int = 20): Array[Int] sampled time serie of number of zombies pursuing humans
• def humansGoneDynamic(by: Int = 20): Array[Int] sampled time serie of number of humans who went out of the world
• def zombiesGoneDynamic(by: Int = 20): Array[Int]) sampled time serie of number of zombies who went out of the world

Global indicators

• totalZombified: Int total number of zombified humans over the course of the simulation
• halfZombified: Int time at which half of humans are zombified
• peakTimeZombified(window: Int = 20): Int time at which the zombification is the most intense (computed by smoothing the time serie with a moving average on a window of size window, and taking the mode of the smoothed time serie)
• peakSizeZombified(window: Int = 20): Int number of zombification when zombification is the most intense (smoothed over a window size window )
• totalRescued: Int total number of humans rescued
• halfTimeRescued: Int time at which half of the humans have been rescued
• peakTimeRescued(window: Int = 20): Int time at which rescue is the most intense (smoothed over a window size window )
• peakSizeRescued(window: Int = 20): Int number of rescue at the time of peakTimeRescued

Spatial indicators

• spatialMoranZombified: Double spatial autocorrelation of the location of zombification events cumulated over time. Take values between -1 (strongest negative autocorrelation) 0 (no spatial autocorrelation) and 1 (strongest autocorrelation)
• spatialDistanceMeanZombified: Double average distance between zombification events
• spatialEntropyZombified: Double entropy of zombification events, or how zombification is uniformally distributed across cells ($\in [0;1]$)
• spatialSlopeZombified: Double level of aggregation of zombification events, can be interpreted as "clustring" intensity.